IP Library Granted Patent US 10,075,193
Granted Patent B2
US 10,075,193 · App. 14/930,879 · Granted Sep 11, 2018

Methods and systems for decoding polar codes

Inventors: Warren Gross (Cote-St-Luc, CA); Gabi Sarkis (San Diego, CA)
Assignee: THE ROYAL INSTITUTION FOR THE ADVANCEMENT OF LEARNING/MCGILL UNIVERSITY
H03M13/3927G06F17/10G06F17/142G06F17/16G06F17/30958G06F17/30961H03M13/1111H03M13/1191H03M13/13H03M13/157H03M13/1575H03M13/617
View Patent ↗
Loading inventors, assignments & file history…
Monitor This Case
Get email alerts when status or documents change.
Order Certified Copies
Most orders are placed with the USPTO same day — all within 24 business hours.
Order via The Patent Place →
Pre-filled with this patent's details
Quick Facts
Patent No.
US 10,075,193
App. No.
14/930,879
Granted
Sep 11, 2018
Kind
B2
Abstract

Herein provided are methods and systems for decoding polar codes. A data flow graph relating to a predetermined polar code is converted to a tree graph comprising rate-zero nodes, rate- 1 nodes, and rate-R nodes. A rate-R node within the binary tree is replaced with a maximum likelihood node when predetermined conditions are met thereby replacing a sub-tree of the tree graph with a single maximum likelihood node.

Claims (30)

1. A method of decoding comprising:

converting a data flow graph relating to a predetermined polar code to a tree graph comprising rate-zero nodes, rate-1 nodes, and rate-R nodes, each of the nodes linked to a root of the tree graph, the tree graph stored in a dual-port random-access memory (RAM);

replacing at least one rate-R node, associated with a sub-tree of the tree graph, with a single maximum likelihood node when predetermined conditions are met, by storing the single maximum likelihood node in the dual-port RAM; and

decoding the tree graph by sequentially visiting each of the nodes at the root.

2. The method according to claim 1 , wherein

the tree graph is associated with a decoder implementing at least one of a pipelined successive cancellation architecture, a line successive cancellation architecture, a semi-parallel successive cancellation architecture, a simplified successive cancellation architecture, and a resource constrained successive cancellation architecture.

3. The method according to claim 1 , wherein

the maximum likelihood node performs a resource constrained exhaustive search maximum likelihood decoding of a constituent code corresponding to a rate-R node it replaces.

4. The method according to claim 1 , wherein

the tree graph is a binary tree.

5. The method according to claim 1 , wherein

the predetermined condition is that (2 k v +1)(n v −1)≤P, where

P is the number of processing elements within a decoder executing the method;

2 k v is the number of candidate codewords; and

n v is the length of the constituent codeword being decoded.

6. A polar code decoder comprising;

a plurality of memory elements, comprising a dual-port random-access memory (RAM), storing a tree graph generated from a data flow graph relating to a polar code being decoded, the tree graph comprising rate-zero nodes, rate-1 nodes, and rate-R nodes each linked to a root of the tree graph; and

a plurality of processing elements configured for decoding the tree graph by:

generating the tree graph from the data flow graph;

replacing at least one rate-R node, associated with a sub-tree of the tree graph, with a single maximum likelihood node when predetermined conditions are met, by storing the single maximum likelihood node in the plurality of memory elements; and

sequentially visiting each of the nodes to decode each of the nodes at the root.

7. The polar code decoder according to claim 6 , wherein

each decoder time step relating to the tree graph is associated with a clock cycle of a clock driving the polar code decoder.

8. The polar code decoder according to claim 6 , wherein

the single maximum likelihood node decodes a sub-tree of the tree graph in a single clock cycle.

9. The polar code decoder according to claim 6 , wherein

the predetermined condition is that (2 k v +1)(n v −1)≤P, where

P is the number of processing elements in the plurality of processing elements;

2 k v is the number of candidate codewords; and

n v is the length of the constituent codeword being decoded.

Assignments (4)
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Jun 6, 2023
From: THE ROYAL INSTITUTION FOR THE ADVANCEMENT OF LEARNING/MCGILL UNIVERSITY
To: 14511581 CANADA INC.
Reel/Frame 063865/0872 →
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Jun 6, 2023
From: 14511581 CANADA INC.
To: 14511581 CANADA LLC
Reel/Frame 063866/0121 →
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Jun 6, 2023
From: 14511581 CANADA LLC
To: POLAR TECHNOLOGIES LLC
Reel/Frame 063866/0334 →
ASSIGNMENT OF ASSIGNOR'S INTEREST Recorded Aug 8, 2017
From: GROSS, WARREN; SARKIS, GABI
To: THE ROYAL INSTITUTION FOR THE ADVANCEMENT OF LEARNING / MCGILL UNIVERSITY
Reel/Frame 043232/0812 →
Continuity (4)
Division 13671617 · Nov 8, 2012
Provisional Application 61556862 · Nov 8, 2011
Provisional Application 61639150 · Apr 27, 2012
Related Publication 20160056843A1 · Feb 25, 2016
Cited By (4)
US 50,983 US 51,025 US 51,036 US 51,037